Project #1: Late-Career Research Productivity and the Timing of Retirement Scientific innovation is playing an increasingly important economic role in the US, and biomedical science in particular generates large health benefits. However, the US scientific workforce is aging rapidly, making it increasingly important to understand how innovative output varies by age, particularly as scientists approach retirement. Existing work has generally focused on the early career. This project will expand our understanding of the age-creativity relationship by studying how innovative output changes during the late career and how the approach of retirement affects innovation. We will begin by (i) documenting trends in innovation (as measured by scientific publications, patents, and citations) in the late career and (ii) estimating how the aging of our innovative workforce will affect the rate of innovation. We will go on to estimate (iii) how the expected age of retirement affects innovation and (iv) how changes in retirement policies will affect the production of innovation. We will use our estimates to simulate how possible changes in retirement policies (such as the movement from Defined Benefit to Defined Contribution pension plans) will affect when researchers retire and the production of innovation. Researchers whose productivity is declining may have an incentive to retire early, so to address the possibility of reverse causation we will use variation in retirement age arising from features of retirement programs such as sharp discontinuities at particular ages or number of years of service and differences in the features of retirement programs across universities) as well as abrupt changes in government funding for scientific research (e.g. stimulus spending). RELEVANCE (See instructions): The US scientific workforce is aging at the very time when the demand for health services is surging with the aging of the baby-boom cohort. This project will inform policy with research-backed predictions for how changes in retirement programs and government science spending will affect the production of biomedical research in the future.